338 REPORTS ON THE STATE OF SCIENCE, ETC. 
Further evidence in support of this large value was found in the behaviour of the 
same mild steel under uniform bending, and also under torsion. By assuming that 
after yield the outer layers only sustain a uniform stress less than that required to 
produce yield, and that the inner layers are still elastic, it was found possible to reduce 
to a very simple form the relation between the applied moment and the deflection, 
or twist in the case of torsion. The curves, corresponding to various values of the 
drop of stress, were plotted, and it was found that in the case of bending the experimental 
points agreed fairly well with the curve for the assumption of a drop of about 25 per 
cent. In the case of torsion the formula showed that with the drop of 25 per cent. the 
moment which the specimen could sustain was exactly equal to that required to initiate 
the yield. In the experiment it was found that this condition was nearly realised, for 
when the yield point was reached the angular displacement increased at a slow but 
uniform rate under the same moment. For the particular steel tested, therefore, the 
16 
ESS (TONS PER SQIN.) 
o 100 200 300 400 500 
yield penetrated right through the specimen, and was not simply confined to a few 
outer layers. Since these experiments were carried out the author has investigated 
the same phenomenon in compression. The results are described in detail in a paper 
on the ‘Strength of Struts,’ published by the Institution of Civil Engineers. Briefly, 
it may be stated that wrought iron, mild steel, and a 36-ton structural steel all show 
a pronounced drop of stress at yield in compression, but the amount is not so great 
as in tension; i.e. wrought iron 7-3 per cent., mild steel 7:5 to 22 per cent., 36-ton 
structural steel 5-9 percent. Bright rolled steel, however, shows no real pronounced 
drop, the stress diagram merely changes its slope very considerably at the point that 
would usually be regarded as the yield, i.e. the diagram is like that of an overstrained 
specimen. The material, however, when annealed does show the characteristic drop 
at yield of mild steel. 
It is the object of this note to consider briefly the bearing of this phenomenon on 
several cases of stress distribution. 
Strength of Struts of Ductile Material. 
If a strut is axially loaded, and the load across the section is uniform, itis clear that 
if any portion of the section yields, and there is a reduction of stress, the resultant of the 
stress no longer passes through the centre of the specimen. There is, therefore, a 
bending moment introduced which still further intensifies the unequal stress distribu- 
tion produced by the yield, and complete collapse is inevitable. In struts of commer- 
